recipe-plan

Pass

Audited by Gen Agent Trust Hub on Sep 23, 2026

Risk Level: SAFEINDIRECT_PROMPT_INJECTION
Full Analysis
  • [INDIRECT_PROMPT_INJECTION]: The skill ingests external data (Design Documents and PRDs) and interpolates this content into prompts for subagents like work-planner and document-reviewer. This creates a surface for indirect prompt injection attacks where a malicious document could influence agent behavior.
  • Ingestion points: Content is read from $ARGUMENTS or files located in docs/design/ and PRD paths.
  • Boundary markers: The skill uses mechanical extraction for prompt building but lacks explicit delimiters or instructions for the agent to ignore embedded commands within the source documents.
  • Capability inventory: The skill facilitates the generation of work plans and test skeletons via tool calls to specialized subagents. It does not perform direct shell execution or network operations.
  • Sanitization: There is no evidence of content sanitization or validation for the ingested documents.
Audit Metadata
Risk Level
SAFE
Analyzed
Sep 23, 2026, 10:09 AM
Security Audit — agent-trust-hub — recipe-plan